Neurobiology of Aging 30 (2009) 11141124

Brain structure and function related to cognormal aging, mild cognitive impairmentCr Carm

Lorena c,b, AJoseersitat

b Institut dInvestigacions Biome`diques August Pi i Sunyer (IDIBAPS), Barcelona, Spainc Alzheimers Disease and Other Cognitive Disorders Unit, Neurology Service, Hospital Clinic de Barcelona, Barcelona, Spain

d Radiology Service, Hospital Clinic de Barcelona, Spaine Departament de Metodologia de les Cie`ncies del Comportament, Universitat de Barcelona, Barcelona, Spain

Received 29 June 2007; received in revised form 7 October 2007; accepted 13 October 2007

Abstract

Cognitiveconsiders heSixteen healinvestigate tinferred premCR measurehealthy eldecerebral netlatter, indicasuperior temCR and actiCR proxiesfunction and 2007 Else

Keywords: CAlzheimers d

1. Introdu

Cognitivity of an ato minimiz

CorresponClnica, Facu08036 Barcel

E-mail ad

0197-4580/$doi:10.1016/jAvailable online 28 November 2007

reserve (CR) is the brains capacity to cope with cerebral damage to minimize clinical manifestations. The passive modelad or brain measures as anatomical substrates of CR, whereas the active model emphasizes the use of brain networks effectively.thy subjects, 12 amnestic mild cognitive impairment (MCI) and 16 cases with mild Alzheimers disease (AD) were included tohe relationships between proxies of CR and cerebral measures considered in the passive and active models. CR proxies were

orbid IQ (WAIS Vocabulary test), educationoccupation, a questionnaire of intellectual and social activities and a composite. MRI-derived whole-brain volumes and brain activity by functional MRI during a visual encoding task were obtained. Amongrs, higher CR was related to larger brains and reduced activity during cognitive processing, suggesting more effective use ofworks. In contrast, higher CR was associated with reduced brain volumes in MCI and AD and increased brain function in theting more advanced neuropathology but that active compensatory mechanisms are still at work in higher CR patients. The rightporal gyrus (BA 22) and the left superior parietal lobe (BA 7) showed greatest significant differences in direction of slope withvation between controls and AD cases. Finally, a regression analysis revealed that fMRI patterns were more closely related tothan brain volumes. Overall, inverse relationships for healthy and pathological aging groups emerged between brain structure and

CR variables.vier Inc. All rights reserved.

ognitive reserve; Brain reserve; Brain volumes; Functional magnetic resonance imaging (fMRI); Cognitive aging; Mild cognitive impairment;isease; Recognition memory; Compensation

ction

e reserve (CR) refers to the hypothesized capac-dult brain to cope with brain pathology in ordere symptomatology (Stern, 2002). CR construct

ding author at: Departament de Psiquiatria i Psicobiologialtat de Medicina, Universitat de Barcelona, Casanova 143,ona, Spain. Tel.: +34 93 4037264; fax: +34 93 4035294.dress: dbartres@ub.edu (D. Bartres-Faz).

was proposed after having observed no direct relationshipsbetween brain damage severity and the clinical manifesta-tion of symptoms. For instance, higher rates of Alzheimersdisease (AD) neuropathology at post mortem examinationswere seen in individuals who were not clinically dementedbut possessed heavier brains and higher counts of largeneurons (Katzman et al., 1988). More recent data is avail-able indicating that a number of participants presentingwith extensive AD-related neuropathology (Braak and Braakstage VI) and/or vascular damage were able to perform at

see front matter 2007 Elsevier Inc. All rights reserved..neurobiolaging.2007.10.008istina Sole-Padulles a, David Bartres-Faz a,b,,Rami c,b, Imma C. Clemente a, Beatriu Bosch

M. Angeles Jurado a, Maite Barrios e,a Departament de Psiquiatria i Psicobiologia Clnica, Univnitive reserve variables inand Alzheimers diseasee Junque a,b, Pere Vendrell a,b,mparo Villar c,b, Nuria Bargallo d,b,Luis Molinuevo c,b

de Barcelona, Barcelona, Spain

C. Sole-Padulles et al. / Neurobiology of Aging 30 (2009) 11141124 1115

a considerably high level in clinical and neuropsycholog-ical examinations, thus showing resistance to the clinicalexpression of neuropathology (Snowdon, 2003).

Two hypproposed tfrom a strutional fashireserve capsuch as theparticular cWhen conBRC woula critical mexceedingwould be udeficits woa more ac

gested by Srelated to tway. Oncewould useticular taskwithin theThe mosteducationaquotient, le(Valenzuel

Direct mal., 1997; CintracraniaTisserand ehave been sreserve. Foreported (Eintracraniabination wicognitive dMortimer e(MCI) insive reportbetween thtion in normTisserandeducation a(Kidron et

Similarlpositron emin the resestudies in tlectual andregional brmainly inregions (Scextendingreports (Alstudies dur

demented individuals found specific brain networks whichwere differentially activated depending on CR background(Scarmeas et al., 2003b, 2004; Stern et al., 2005).

espiteution aD signI) studspectiechniCR. Aetho

intelliewed ipathol; Ernsated t

rebralonly freted

d, Hahe patin heaces inhy elden whinexi

ures a

les ofive evation,hologn issurelaticordi

(Wolfntingim ofs betws of fMhy eldatient

ethod

Subjec

rty-fonformled inhy eldpatiense anServicary callar dotheses based on reserve mechanisms have beeno account for the abovementioned findings, onectural point of view and the other of a more func-on (Stern, 2002). At an anatomical level, the brainacity (BRC) model ascertains that certain factors,number of synapses and brain volume, confer aapacity to endure neuropathological processes.

sidering dementia and pre-dementia conditions,d be able to prolong the preclinical stage untiloment would be reached. From the moment of

that threshold on, vulnerability to brain damagenavoidable and eventually, clinical and functionaluld be evident (Satz, 1993). On the other hand,tive or functional model of CR has been sug-tern (2002). According to this view, CR would be

he ability to recruit brain networks in an effectivepathological processes begin to occur, subjectsalternative networks in order to perform a par-successfully or to maintain ones clinical statusnormality, a process defined as compensation.

frequently used proxies reflecting CR comprisel/occupational attainment, premorbid intelligenceisure, cognitive and mental stimulating activitiesa and Sachdev, 2005).easures of brain (Katzman et al., 1988; Kidron etoffey et al., 1999; Edland et al., 2002), head, or

l size (Schofield et al., 1995; Jenkins et al., 2000;t al., 2001; Mortimer et al., 2003; Wolf et al., 2004)tudied in the aging literature as surrogates of brainr instance, although negative findings have beendland et al., 2002; Jenkins et al., 2000), reduced

l volume or smaller head size by its own or in com-th low education may confer an increased risk forecline and dementia (Schofield et al., 1995, 1997;t al., 2003) including mild cognitive impairment

old age (Wolf et al., 2004). Further, inconclu-s have been published regarding the relationshipese variables and proxies of CR such as educa-al aging (Coffey et al., 1999; Edland et al., 2002;

et al., 2001) with inverse relationships betweennd brain volumes in the case of established AD

al., 1997).y, at a functional level, investigations mainly usingission tomography (PET) have been undertaken

arch of CR in healthy elders and AD. At resthe latter condition found that education and intel-social life activities were inversely correlated withain metabolic activity and/or cerebral blood flowtemporal, parieto-temporal and parieto-occipitalarmeas et al., 2003a; Perneczky et al., 2006) alsoto other cortical and subcortical areas in otherexander et al., 1997). Further, reports of activationing cognitive tasks among demented and non-

DresolBOL(fMRity, rethis tgatethis mfluid(revitinct2006roborof ceever,

interpregarthat ttaskferenhealt

Evtle ormeas

samppositeducmorpity, athesefer acdatapresethe alationmean

healtAD p

2. M

2.1.

Foten ienrolhealtADdiseaogyprimCasteits lower invasiveness, higher spatial and temporalnd the knowledge that increases and decreases ofal from functional magnetic resonance imagingies reflect increases and decreases in neural activ-vely (Logothetis et al., 2001; Shmuel et al., 2006);que has been less frequently applied to investi-

number of reports have been published usingdology to indicate that measures such as generalgence are related to variations in BOLD activityn Jung and Haier, 2007). Further, studies in dis-ogical conditions (Chang et al., 2006; Cader et al.,t et al., 2002; Bartres-Faz et al., 2006) have cor-he capacity of this technique to reveal the usagereserve mechanisms during cognitive tasks. How-ew investigations were specifically designed andin terms of the cognitive reserve theory. In this

beck et al. (2003) and Stern et al. (2003) foundtern of activation during a nonverbal recognitionlthy young subjects was related to individual dif-CR variables. Yet, very scarce data is available iners or among AD patients.en considering previous findings, there is still lit-stent data of how functional and structural brainre distinctly related to CR measures in the sameindividuals. Since previous reports have found

idences for a relationship between accounts ofoccupation, or premorbid IQ variables and bothic brain measurements and patterns of brain activ-e of interest would be to determine the effects ofonships in the same sample of individuals that dif-ng to their clinical status. Specifically, very scarceet al., 2004) addressed these questions in patientshigh-risk conditions for AD such as MCI. Thus,the present report was to investigate the corre-een the main proxies for CR, brain activity (byRI) and cerebral structural characteristics among

ers, patients diagnosed as having MCI and milds.

ts

ur subjects older than 65, who provided writ-ed consent (or their relatives in AD cases) were

the study. The whole sample comprised 16ers, 12 amnestic MCI (MCI) cases and 16 mildts. Participants were selected from Alzheimersd other cognitive disorders unit, at the Neurol-e, Hospital Clinic of Barcelona, and from are health centre in the area of Barcelona (CAPel Valle`s). All subjects underwent clinical and

1116 C. Sole-Padulles et al. / Neurobiology of Aging 30 (2009) 11141124

neuropsychological evaluations. The diagnostic proceduresemployed to classify individuals into the abovementionedgroups have been described elsewhere (Rami et al., 2007).Briefly, heatia, and diddid not exsecondaryneuropsychsis and abMCI was dal. criteriaLopez et acal judgme(2) impairepreserved gdaily livingment had1.5S.D. beage and edor medicalA previousText Memoas an episooff below tlevel. Probciplinary clone neuropteria werefunctionalincluded wAtypical Aimpairmen

2.2. Proxie

Three mCR literatuAdult Intelsubtest, ad(Lezak et aeducationeducationacoded as innal valuesschool, 2 =sity educatmanual, 1technician,4 = managevalue was ovalues (ranto consider2007) incluand cognitiplaying, paing) and so

associations, voluntary work). These measures were gatheredinto a customized questionnaire with scores ranging from0 to 19, the greater the score indicating increased CR. The

tionnag in tof patly, tobles, asing f

wing te factoe com

Magn

ll 44exam

(Geneightetion-dimeions i[TE] == 1, Fes w

s withired us

2000went

a slice

Struct

atistic.5 wa

e invee preption ach waementasure

ddingwas

y + whin oes pr

chofi, 2004id notthese vtigatinrity (ooveralthy individuals did not meet criteria for demen-not present cognitive complaints. Further, they

hibit cognitive performance below 1.5S.D. in amemory test or in any other test comprised inological examinations of language, praxis, gno-

stract reasoning (Rami et al., 2007). Amnesticiagnosed according to the modified Petersen et(2001) and two additional criteria, similar to

l. (2003): (1) memory decline according to clini-nt and preferably corroborated by an informant,d memory function for age and education, (3)eneral cognitive function, (4) intact activities ofand (5) non-demented; (6) the memory impair-

to be of the episodic memory type defined bylow the control group mean, taking into accountucational level, and (7) absence of psychiatriccauses accounting for these memory problems.

ly validated normative Spanish test: the Delayedry Test (Pena-Casanova et al., 1997) was useddic memory test for determining a 1.5S.D. cut-

he mean, taking into account age and educationalable AD diagnosis was established by an interdis-inical committee formed by two neurologists andsychologist. DSM-IV and NINCDS-ADRDA cri-applied taking into account clinical and objectiveand neuropsychological results. All AD patientsere mild AD (Global Deterioration Scale-4 stage).D variants with non-significant episodic memoryt were excluded from the study.

s of cognitive reserve

ain proxies reflecting those commonly used in there were defined. The first one was the Wechslerligence Scale 3rd version (WAIS-III) Vocabularyministered as a measure reflecting premorbid IQl., 2004). A second CR variable was defined asoccupation and included quantifications of both

l and occupational attainment. This measure wasa previous study (Staff et al., 2004) using ordi-

as follows: 0 = no formal education, 1 = primarysecondary education and 3 = superior or univer-ion and as regards occupation; 0 = non-qualified= qualified manual, 2 = qualified non-manual or

3 = professional (university degree required),r or director (university degree required). The finalbtained by adding the education and occupation

ge 07). A third proxy taken into account as an aimother relevant variables related to CR (Scarmeas,ding recordings of lifetime occupations in leisurevely stimulating activities (reading, writing, musicinting) as well as physical (sports and daily walk-cial life (participation in social activities or groups,

quesipatincase

Finalvariaject ufollosinglof th

2.3.

AMRI1.5TT1-wtificathreenicattimetionsvolumsliceacqu(TR =90). Twith

2.4.

Stlab 6singlimagmina(whiimpla me

by asure

([graumes

volumtus (Set al.we dwithinvesintegtheirire was administered directly to the subject partic-he study with the presence of their relatives in theients to ensure the validity of the data provided.summarize the information relating the three CRcomposite CR score was obtained for each sub-

actorial analyses (principal component methods)he procedure described by Stern et al. (2005). Ther extracted (composite CR) accounted for 79.6%

mon variance of these three measures.

etic resonance imaging acquisition

subjects underwent structural and functionalinations. Scans were obtained on a GE Signaeral Electric, Milwaukee, WI). High-resolutiond images were acquired for anatomical iden-

with a Fast Spoiled Gradient-Recalled Echonsional sequence (Digital Imaging and Commu-n Medicine) format: repetition time [TR]/echo

12/5.2, inversion time = 300, number of exita-OV = 24 3 24 cm, 256 3 256 matrix. Whole-brainere acquired in an axial plane yielding contiguousslice thickness of 1.5 mm. Functional images wereing a T2*-weighted gradient echo planar imagingms, TE = 40 ms, FOV = 24 3 24 cm, flip angle of

y axial slices were obtained for each brain volumethickness of 5 mm and a gap of 1.5 mm.

ural MRI

al Parametric Mapping (SPM2) running in Mat-s used to analyze structural brain images. Astigator performed the prior manual steps inaration (anteriorposterior commissure line deter-nd image reorienting). Following segmentations performed against the T1 template (MNI)

ed in SPM) of the three tissue compartments,of whole-brain volume was obtained in mm3the gray and white matter volumes. This mea-then corrected for whole intracranial volumeite]/[gray + white + CSF]). We used brain vol-ur study instead of head size or intracranialeviously employed to reflect premorbid brain sta-eld et al., 1995, 1997; Mortimer et al., 2003; Wolf; Edland et al., 2002; Jenkins et al., 2000) becauseaim to determine a risk for AD or MCI associatedariables but instead we were interested in directlyg the correlations between CR proxies and brainr otherwise atrophy) for each subject relative to

ll head size.

fanavaranHighlight

fanavaranHighlight

C. Sole-Padulles et al. / Neurobiology of Aging 30 (2009) 11141124 1117

2.5. fMRI procedure and memory recognitionassessment

Originamat GE-adorganizedsubject) bNottinghampatible witcommissurthe scans twere transfmalized imkernel (fulexperimentrepeatedeach) presetion of thereached. Tuli period 1consisted osame picturthis conditiblock (presregards the50 non-emwere told bshowing adifferent actal block,scanning, sthey were gdecide whisession. Th50.

2.6. Data a

The Statwas used tclinical antests whenage, gendeCR measugroup. Furtest whetheumes diffeSPSS wereanalyze fMyses for eactivity obpared withexperimentwithin-grousimilar maCR). Withi Tab

le1

Dem

ogra

phic

glob

alco

gniti

ve

and

CRvar

iabl

esac

ross

the

thre

est

udie

dgr

oups

Cont

rols

(n=

16)

MCI

(n=

12)

Mild

AD

(n=

16)

F/

2Po

stho

c(S

cheff

e)A

ge73

.31

(4.90

)[CI

70.7

75.

9]74

.25

(6.18

)[CI

70.3

78.

2]76

.50

(5.80

)[CI

73.7

76.

4]1.

36n

.s.

Gen

der(

M/F)

a5/

112/

105/

110.

94n

.s.

MM

SE27

.75

(0.86

)[CI

27.2

928

.21]

25.8

3(2.

18)[

CI24

.45

27.2

2]21

.69

(4.03

)[CI

19.5

423

.83]

20.3

8C

>A

D**

*;M

CI>

AD

***

Rec

ogni

tion

mem

ory

(fMRI

)43

.37

(5.21

)[CI

40.6

46.

2]36

.25

(6.88

)[CI

31.9

38.

1]26

.31

(5.35

)[CI

32.4

38.

1]35

.43

C>

MCI

**;C

>A

D**

*;M

CI>

AD

***

Cogn

itive

rese

rve

(CR)

Com

posit

eCR

0.66

(0.80

)[CI

0.23

1.0

9]0

.40

(0.77

)[CI

0.8

9to

0.09

]0

.36

(1.03

)[CI

0.9

1to

0.19

]6.

91C

>M

CI*;

C>

AD

**Vo

cabu

lary

WA

IS-II

I43

.56

(9.06

)[CI

38.7

48.

4]30

.08

(7.99

)[CI

25.0

35.

2]29

.75

(14.54

)[22

.037

.5]7.

64C

>M

CI*;

C>

AD

**Ed

ucat

ion

occu

patio

n4.

56(1.

86)[

CI3.

15.

6]2.

42(2.

02)[

CI1.

63.

7]2.

75(2.

21)[

CI1.

63.

9]4.

79C

>M

CI*

CRqu

estio

nnai

re9.

00(4.

24)[

CI6.

711

.3]

5.67

(3.82

)[CI

3.2

8.1]

5.56

(3.93

)[CI

3.5

7.7]

3.63

n.s

.

Valu

esar

egi

ven

inm

ean

s(st

anda

rdde

via

tions

);n

.s.,

no

n-s

igni

fican

t;*p